st, ed = ed, ed+FLAGS.batch_size
times += 1
loss /= times
acc /= times
return acc, loss
def inference(model, sess, X): # Test Process
return sess.run([model.pred_val], {model.x_: X})[0]
with tf.Session() as sess:
if not os.path.exists(FLAGS.train_dir):
os.mkdir(FLAGS.train_dir)
if FLAGS.is_train:
X_train, X_test, y_train, y_test = load_cifar_2d(FLAGS.data_dir)
X_val, y_val = X_train[40000:], y_train[40000:]
X_train, y_train = X_train[:40000], y_train[:40000]
mlp_model = Model()
writer = tf.summary.FileWriter(FLAGS.train_dir)
writer.add_graph(tf.get_default_graph())
writer.flush()
if tf.train.get_checkpoint_state(FLAGS.train_dir):
mlp_model.saver.restore(sess, tf.train.latest_checkpoint(FLAGS.train_dir))
else:
tf.global_variables_initializer().run()
pre_losses = [1e18] * 3
best_val_acc = 0.0
acc /= times
return acc, loss
def inference(model, X): # Test Process
model.eval()
pred_ = model(torch.from_numpy(X).to(device))
return pred_.cpu().data.numpy()
if __name__ == '__main__':
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
if not os.path.exists(args.train_dir):
os.mkdir(args.train_dir)
if args.is_train:
X_train, X_test, y_train, y_test = load_cifar_2d(args.data_dir)
X_val, y_val = X_train[40000:], y_train[40000:]
X_train, y_train = X_train[:40000], y_train[:40000]
mlp_model = Model(batch_norm=args.batch_norm, drop_rate=args.drop_rate)
mlp_model.to(device)
print(mlp_model)
optimizer = optim.Adam(mlp_model.parameters(), lr=args.learning_rate)
# model_path = os.path.join(args.train_dir, 'checkpoint_%d.pth.tar' % args.inference_version)
# if os.path.exists(model_path):
# mlp_model = torch.load(model_path)
pre_losses = [1e18] * 3
best_val_acc = 0.0
epochs = []
train_data = {"loss": [], "acc": []}